Self-Driving Car Sim

You set three numbers — proportional, integral, and derivative gains — and then watch a small car attempt to follow a winding 2D track on its own. The car uses a PID controller, the same family of algorithms that steers real autonomous vehicles, industrial robots, and cruise-control systems. Nudge the parameters and the car might hug every curve cleanly; push them too far and it wobbles, overshoots, or spirals off the road entirely.

That cause-and-effect loop is the whole point. There's no score, no timer — just the satisfaction of tuning a feedback loop until the steering feels right. It's a surprisingly tactile way to build intuition for control theory without reading a single equation. If you enjoy hands-on systems like this, browse more interactive simulations in the catalog, or try Ant Colony to see another algorithm learn its environment in real time.

Works best on a desktop where you can fine-tune sliders precisely; a mouse gives you much more control than a touchscreen here.